1. Introduction
Nowadays, the quality of life depends mainly on electrical devices, shaping how people dwell, work, recreate, and transport. According to [
1], of the energy consumed worldwide, 21.69% is by residential buildings, 18.22% by the commercial sector, 27.84% by the transportation sector, and the rest by the industrial sector. Nevertheless, the use of energy is compromised mostly by how this energy comes, which is from thermoelectric plants that generate carbon dioxide emissions that threaten the quality of life from a global perspective. Therefore, it is essential to use energy efficiently and include renewable energy sources, which cannot replace the energy from thermoelectric plants. The level of technology reached today allows monitoring, measuring, controlling, and scheduling electrical appliances or devices in real time at home, work, and public places [
2]. At home, modern electrical devices allow people to have the comfort level demanded today, facilitating domestic tasks, home office, homeschooling, recreational activities, entertainment, and the involvement with the community to which they belong [
3].
Currently, energy manager systems (EMSs) allow handling the energy consumed by a group of people in a household and provide specific tools to make it as efficient as possible [
4]. Nevertheless, the variety of electrical devices and their complexity make their integration, programming, or effective use within the EMS difficult, especially for those not related to cutting-edge technology, like senior users or digitally illiterate users [
5]. Moreover, people develop energy consumption patterns that are usually related to cultural or psychological outcomes that are difficult to change [
6]. Usually, residential consumers do not have the tools to measure and alter their energy consumption or control electrical devices when energy consumption is metered only monthly [
7]. New tools and methodologies to improve the estimation accuracy of residential energy consumption are necessary to improve energy-saving potential calculations [
8]. In this way, modifying habits or the creation of new ones may be possible, overlooking energy demand in the market and price variations, or to allow consumers to understand their environmental impact related to electricity consumption.
Gamification strategies allow people to change their consumption and social involvement through incentives, environmental awareness, and possible competition and cooperation with other community members in similar conditions [
9]. Gamification techniques applied to EMS make it possible to stimulate users to diminish their energy consumption and save money on billing [
10], which entails reducing greenhouse gas emissions produced by the primary electricity grid. In addition, these techniques may favor renewable energies and allow for exchange of information in real-time between suppliers and consumers for energy resilience and security. They provide valuable tools for the energy market to better distributor system operators (DSOs) and demand response (DR) programs while creating or increasing the community’s sense of belonging [
11].
Current approaches for gamified saving-energy strategies try to positively influence the behavior of the users towards efficient consumption by socio-technical systems, which demonstrated that managing the consumers’ demand gives a more sustainable consumption [
12]. One of these applications is Wattsup [
13], that displays energy consumption and CO
2 emission data through a social media application, giving users the ability to share and compare household data with their friend. This app uses an energy monitoring device which transmit the data to a server for a a social media gamified application. Another interesting project is enCOMPASS [
14], in which a gamified web application was developed that is accessible via PC and cell phone, enabling interactive visualization of energy consumption patterns.
Using current technologies in artificial intelligence as an adaptive neuro-fuzzy inference system (ANFIS), fuzzy logic, or neural network decision systems gives insights regarding the type of gamification elements that can be displayed within the human–machine interface (HMI) environment to promote electrical energy reduction [
15]. The relevance of adaption of these artificial techniques is that they emulate human making decisions so that reliable proposals can promote energy reduction. In this way, it is possible to think of an integrated and complex system for efficient energy management that favors renewable energies, creates awareness of consumption and energy savings, and provides analysis tools to improve techniques and algorithms related to forecasting consumption patterns. This project uses AI and fuzzy inference to recognize consumption patterns and calculate potential changes in users’ behavior to achieve energy efficiency, while offering an uncomplicated and custom interface to the user. Current gamified approaches do not offer a tailored interface for user engagement.
Therefore, this paper presents three types of users based on their preferences and goals. Then, it analyzes the energy usage impact for each home located at Concord, California, with a focus on the heater/furnace and air conditioner. Finally, this proposal presents a tailored gamified application linked to an EMS for each case and the proposal of flexible loads required during on-peak periods and the time of use (ToU). This gamified application uses an ANFIS decision system where the inputs consider the electrical consumption and the set point during the summer or winter seasons to deliver the type of gamified motivation needed to promote household appliances’ flexible usage.
The structure of this paper is as follows:
Section 2 presents the state of the art regarding EMS, gamification, ANFIS, and thermal comfort.
Section 3 presents the three-step methodology used for this paper.
Section 4 describes the results, by step, of the previous section and the three tailored gamified interfaces for each case.
Section 5 presents a discussion regarding the current EMS and the gamification approach with the advantages and disadvantages of the present proposal. Finally,
Section 6 details the conclusions and future work.
2. State of the Art
Today’s technologies aim to inter-connectivity, automation, and high performance of the electrical grid using the Internet of Things (IoT) and artificial intelligence (AI) technologies such as big data (BD) and machine learning (ML) [
16,
17]. The smart new paradigm in the electrical grid allows the energy provider and end-user to track the energy consumption in real time and know how energy is being consumed in each electrical device, known as energy disaggregation and usually done through non-intrusive load monitoring (NILM) techniques [
18,
19]. This provides the opportunity to solve particular problems such as energy efficiency through the existing tools (e.g., smart metering infrastructure, smart electrical devices, smart plugs and sensors, internet, programming tools, and user interfaces) in order to build the smart grid and control the electrical devices involved [
20,
21]. A well known tool to deal with energy efficiency is EMS [
22], where the information related to energy consumption, the electrical market, the user preferences and consumption patterns, as well as indoor thermal comfort and outdoors or environmental conditions can be merged into a decision-making process for optimizing energy usage.
2.1. Energy Management System
EMS is a computer-aided system to monitor, control, and optimize the generation, distribution, and consumption of the electricity within the grid, keeping the balance between supplied and demanded energy at any given time, managing the available DERs, the load scheme, and energy exchange with the primary grid [
23]. EMS presents information about the electrical network status (e.g., energy stored, forecast energy production by distributed generators (DG), appliances scheduling), enabling decision-making about its safe and cost-effective operation [
24]. Likewise, EMS would collect generation, consumption, and storage information of past and current performance to improve the decision-making process, optimal manipulation of controllable devices, consumption, and generation forecasting and, finally, network management recommendations. It would also provide relevant information on the weather, the energy market, and billing user status [
2]. In this way, EMS would manage controllable loads using communication technologies, sensors, and actuator devices, nowadays usually included in electric devices or modern appliances, thus, enhancing the cost-effective and reliable operation of the user electrical grid, a smart home in this case, by active participation in the electricity market [
25].
The current trend favors individualized and private monitoring of energy resources, facilitating the inclusion of distributed energy resources (DER) such as low voltage generators from renewable sources, electric vehicles (EV), and optimal managing of programmable and controllable devices such as thermostat HVAC systems [
26,
27,
28]. In [
27], uncertainty and load demand variability in a smart home are analyzed without the user’s preferences nor goals, and in [
29], a certain level of comfort is assumed. Simulation frameworks [
30,
31] control electrical devices in a dynamic price scheme but do not consider human behavior as a part of the equation to achieve energy efficiency. In this way, the social part needs to be seen so the end-users can adopt the EMS without negatively affecting their social behavior, where consumption patterns and DR programs are allowed to reach energy efficiency and then achieve a smart and sustainable electrical grid that is required by society [
32].
Nowadays, people can reduce and manage the electricity consumption in homes by installing home energy management systems (HEMS). Information and communication technologies (ICT) will enable two-way communication among the customers and distributors, providing real-time rates and billing status [
2]. The EMS requires a user-friendly interface, energy consumption display, auto-configuration, or easy set-up to enhance user interaction with its energy distributor. Users will be able to optimize consumption when the price is high, and distributors will be able to shift and shape demand, providing statistical data to predict energy consumption. In this way, EMS allows the generation of databases of consumption patterns used in algorithms to optimize energy consumption [
33]. Consumption patterns are essential to predict energy consumption, shape the energy demand, and favor renewable sources. Energy consumption patterns of a household contain the load scheme of appliances and electrical devices as EVs. The historical performance of DERs, such as photovoltaic panels, wind turbines, and batteries, are used to propose to the user to change their consumption habits by scheduling appliances and suggesting strategies. [
34].
Therefore, the human factor must be included in the electrical simulators using probability functions based on actual data. In addition, one way to emulate the users’ response on DR programs and their interaction with DSOs is that gamification strategies can be used to study consumption patterns and how to change them to achieve energy efficiency. This is possible modeling electrical cases through a network of interconnected agents in order to test stochastic behaviors.
Figure 1 shows the connected energy entities and the interaction among the load scheme, end-user, and the energy provider.
2.2. Gamification and Serious Games
Gamification uses game elements and game design techniques in a real-life context [
9]. Some gamification strategies propose modifying the demand by taking advantage of the energy generated by DERs that usually work when demand is much lower than during peak hours [
35]. Moreover, the users compete in shared spaces to save energy in real-time, using challenges, social sharing, rewards, leader-boards, points, tips, levels, rankings, avatars, badges to promote environmental education, consumer awareness and users engagement [
36]. Therefore, gamification’s primary goal is to motivate users by using game-like techniques in the real world to shape individuals’ behavior and improve their skills [
12].
In [
37], the authors analyzed Fit for Green, PowerAgent, Greenify, and PowerHouse gamified applications to suggest three best practices for sustainable applications: make sustainability a fun and rewarding experience, create positive peer pressure sustainability issues, and use gamification to promote meaningful action. Regarding the gamification elements, the applications and their considerations are as follows:
Fit for Green uses feedback and rewards by employing cardio machine workouts to promote environmental awareness through two impact workouts. The first by promoting exercise and feeding that energy back into the grid. The second, by generating funds for charities that protect the environment.
Greenify focuses on motivating senior users to become aware of climate change by enhancing social sharing and tips between them to address this problem.
PowerHouse uses avatars and archetypal characters to promote the sense of belonging of the end-user so the users accept exploring the cause and effect relationships of daily activities regarding electrical consumption and receive instant feedback.
PowerAgent uses appealing computer games and mobile applications to promote changes in household users’ energy consumption patterns.
Moreover, an energy gamified application must be environmentally goal-oriented with game-like features [
38]. In [
36], they analyzed nineteen gamified projects from Europe to propose these game design elements to engage end-users in energy applications: statistics, messages, tips, discounts in electricity bill, virtual currency, prizes, offers and coupons, competition, collaboration, energy community, dashboard, leader-board, progress bar, message box, notifications, degree of control, points, badges, and levels.
Table 1 shows the extrinsic and intrinsic motivations used in this paper [
32,
38].
Extrinsic motivation: People are motivated because they want something they cannot get, and earning it infers outer recognition or even monetary prizes. Include factors of external control, identification, and integration.
Intrinsic motivation: The activity is rewarding on its own without a particular purpose to succeed. This motivation considers autonomy, competence, and relatedness.
As described in [
36], a gamified energy application framework can be compound system of technical, behavioral, and economic systems. The technical component has the smart metering system, EMS, web/mobile applications, network, and software, which are necessary to monitor and control energy consumption and user response. The game design elements for the behavioral aspect are information provisioned, rewarding system, social connection, user interface, and performance status. For the economic aspect or value proposition, the components are the residential customers, suppliers, and society, which bring significant value streams to users while driving positive and measurable business outcomes for energy providers and society.
Previous research includes the use of AI for residential EMS with no prior linkage to a gaming strategy that engages the end-user in the process of energy reduction [
39,
40,
41]. In addition, the previous gamified strategies did not consider personalizing interfaces for energy reduction [
42,
43,
44,
45,
46] or only proposed frameworks with no interface proposals [
47,
48]. Hence, some proposals include incorporating tailored gamified interfaces through a three-step framework that continuously runs through the HMIs to receive updates and feedback and make adjustments to the gamified interface to engage, teach, and motivate end-users to save energy in connected thermostats [
49,
50,
51,
52], smart homes [
53,
54], smart communities, and smart cities [
55].
2.3. Adaptive Neuro-Fuzzy Inference Systems (Anfis) for Gamified Interfaces
A fuzzy inference system is a fuzzy-rule-based system consisting of linguistic rules or conditional states expressed in the form IF A–THEN B, where A and B are labels of fuzzy sets. Fuzzy systems are used to capture human thinking or the reasoning of human ability to make decisions in an environment of uncertainty and imprecision [
56]. On the other hand, an adaptive network is a structure consisting of nodes and directional links through which the nodes are connected, their outputs depend on the parameters about these nodes, and the learning rule specifies how these parameters should be changed to minimize a prescribed error measure [
57]. In that regard, several proposals have been made, such as the combination of artificial neural networks with fuzzy systems. Artificial neural networks can learn and adapt from experience, thus complementing fuzzy systems. Among the most important techniques is ANFIS, an adaptive neuro-fuzzy inference system proposed by Jang [
58] in 1993, which generates fuzzy IF–THEN rule bases and fuzzy membership functions automatically. ANFIS is based on adaptive networks, a super set of feed-forward artificial neural networks with supervised learning capabilities as stated by Jang in [
58] and [
59]. The basic learning rule of adaptive networks is based on the gradient descent and the chain rule; however, this method is usually slow and likely to become trapped in local minimal. Thus, Jang proposed a hybrid learning rule that combines the gradient method and the least-squares estimate to identify parameters.
In [
54], they proposed the inclusion of Alexa and cameras to track the senior people and monitor their daily mood to improve their quality of life by promoting social inclusion and physical exercise. The multi-sensor system is used within a smart home environment to identify the physical characteristics of older people. Thus, the voice and face detection are evaluated on an ANFIS system to propose the personalized gamified elements that run in an HMI needed for each type of user.
In [
55], based on the type of environmental home and the amount of electrical energy usage, they used the ANFIS decision system to propose a gamified interface based on intrinsic or extrinsic motivations.
2.4. Thermal Comfort
Thermal comfort is essential in a built environment for energy saving, where data-driven thermal comfort models enhance the prediction accuracy to maintain optimal the human comfort reaction and its interaction with the environment. The existing thermal comfort models are applied in different environments like sleeping environments, indoor and outdoor environments. These models consider features such as groups or types of people, such as elderly and different races, gender, age, weight, the amount of activity, clothing thermal resistance, air temperature, radiation, relative humidity, wind speed, activity intensity, metabolic rate, and other factors [
60]. Besides physiological aspects, weather conditions, and the level of activity and occupancy in the house, psychological aspects and users’ preferences are important to set up thermal comfort [
60].
The Universal Thermal Climate Index (UTCI) considers a reference environment with 50 percent relative humidity, vapor pressure below 20 hPa, air temperature, and wind speed of 0.5 m/s at 10 m height or 0.3 m/s at 1.1 m. In addition, the thermal stress is categorized within the ten ranges of different values of the UTCI [
61]:
Extreme heat stress: above 46 °C
Very strong heat stress: +38 to +46 °C
Strong heat stress: +32 to +38 °C
Moderate heat stress: +26 to +32 °C
No thermal stress: +9 to +26 °C
Slight cold stress: 0 to +9 °C
Moderate cold stress: 0 to −13 °C
Strong cold stress: −13 to −27 °C
Very strong cold stress: −27 to −40 °C
Extreme cold stress: below −40 °C
Thermostats stand for managing thermal comfort and energy consumption, whether temperature is good enough in the home, and how much comfort users are willing to concede to save energy and money. Because thermal comfort has to do with human psychology, there are many fuzzy elements in the modeling of these systems, where technologies such as machine learning and big data help create adequate and functional models [
62]. ANFIS is widely used in thermal comfort models to calculate building energy needs by controlling humidity and temperature in HVAC systems, and thermostats [
15,
63,
64]. In this study, the buildings’ construction material is taken into account to calculate the indoor temperature with the outdoor one. This information is important when thermal comfort is delimited form users preferences to calculate the energy-saving potential.
4. Results
This section presents the results from the proposed methodology, describes each step’s results, and presents the tailored gamified interfaces based on the type of energy user based on their preferences regarding saving money, comfort level, environmental commitment, and technology knowledge. Thus, three cases are presented: a user who has a green attitude, a user who does not care about saving energy, and a user who wants to make home improvements while saving energy.
4.1. Step 1
Figure 3 shows the user’s preferences goals regarding saving money, comfort level, environmental commitment, and technology knowledge. It is essential to understand and profile the users better such that flexible loads can be proposed based on their needs and expectations during this step.
Using EMS framework simulation, it is possible to study the consumption patterns of an average household considering family characteristics such as the number of people and level of activity using their appliances. The level of comfort can be determined by factors like environmental conditions, house infrastructure, users’ willingness to modify consumption patterns to save money and energy, and environmental commitment.
For the purpose of this paper, three types of user were selected based on their energy awareness and motivation to modify their energy consumption by changing the time of use of the household appliances:
Case 1. Home-focused energy saver—they are concerned about saving energy and interested in home-improvement efforts.
Case 2. Green advocate—show the most positive overall energy-saving behavior, have the most robust positive environmental sense, and high interest in new technologies.
Case 3. Disengaged energy saver—less motivated to save energy through energy savings. They are not concerned about the environment nor new technologies.
4.2. Step 2
July 1st for the summer period and December 16th for the winter period were selected for this analysis. The outdoor temperature was obtained from the Statistic Report of the annual weather file (stat file) [
79].
A typical year for this place involves use of the cooling system from April through October, and from November to March, use of the heating system.
Table 4 describes the selected scenarios that emulated the energy consumption with different thermal conditions in different seasonal times.
Figure 10a depicts the summer period considering the indoor temperature and the setpoint for each case; the same in
Figure 10b for winter season. Case 3 is consistent for an energy waster user as they prefer lower temperatures during summer and higher temperatures during winter. Case 1 is for a home-focused selective energy saver, which compared with case 2 and case 3, is between the home-focused selective energy saver and the disengaged energy saver. This graph shows that the home-focused energy saver (case 2) is oriented in saving energy while improving their home. During winter periods, this user can be motivated to change their consumption patterns by reducing the thermostat setpoint at least 1 °C and therefore saving energy without affecting the thermal comfort at home. For the disengaged energy saver user, as they are not interested in saving energy, the efforts need to be oriented to saving money strategies and into a rewarding system. The indoor temperatures during summer require air conditioners; however, the strategy needs to be oriented more in increasing the setpoint at least 1 °C. The UTCI scale demonstrates that the user could be without thermal stress from 9 to 26 °C [
61]; at lower temperatures it is assumed that the user is wearing warmer clothes; at higher temperatures that they are using less or not heavy clothes. In addition, by increasing the setpoint by 1 °C during summer periods, at least 6% of the electrical bill could be saved [
80].
Table 5 depicts the average daily electrical consumption profile for all the household appliances in the home in the summer and winter period. Therefore, considering the summer period and the kWh of the weekday times 365 days, the results for a year in each case is
Case 1: 24.33 kWh/day × 365 days = 8881 kWh
Case 2: 22.93 kWh/day × 365 days = 8370 kWh
Case 3: 25.43 kWh/day × 365 days = 9282 kWh
This was calculated this way due to more kWh being consumed in the summer period than in the winter period. The weekday was selected as the weekends have atypical consumption, and not every weekend the end-user is spending that electrical energy. Energy consumption does have not appreciable changes during weekdays since users share common zones such as living room and kitchen, but the increase in domestic tasks, such as laundry, on weekends and the usage of appliances such as the refrigerator increase a little bit. The lighting system is for a big house without EMS supervision or sensor care.
Figure 11 and
Figure 12 show the consumption patterns during weekdays and the weekend in summer and winter periods, respectively, with the load scheme due to AC or heater, lighting system, stove, dishwasher, refrigerator, washer machine, dryer, and water pump.
4.3. Steps 3 and 4
The household’s load scheme obtained from the database has the following appliances: air conditioner (AC), furnace/heater, dryer, stove, lighting system, dishwasher, refrigerator, clothes washer, and water pump. The study focuses on thermal comfort and energy savings, and the HVAC system is the most flexible and suitable to modify user consumption behavior. According to the decision tree diagram in
Figure 7, the AC system in summer and the electric furnace or heater in winter have the most flexible range situated in the auto-reduction and auto-reschedule according to the time of the day and level of occupancy.
The ANFIS system has two inputs, the daily electrical consumption in kilowatts [kWh] and the set point temperature. To obtain the daily consumption for a year, the average high 3C IECC climate zone consumption was selected from
Table 3. Then, considering the standard deviation and the mean, 365 values were created, giving an annual consumption of 8894.7 kWh, similar to Case 1. The set point temperature uses the occupied values from the energy saver user; it goes from 15 to 26 °C. The output is related to the type of gamified motivation, intrinsic, extrinsic, or both, and considers the following:
A home that consumes more energy with a set point below 21 °C for cooling and above 20 °C for heating requires extrinsic motivation for outer recognition and external rewards. A home that consumes less energy with a set point above 23 °C for cooling and below 18 °C for heating can be related to intrinsic motivation. The house uses less kWh than the other in similar conditions. On the other hand, the average home and set point below 23 °C and above 21 °C for cooling and set point below 20 °C and above 18 °C have both motivations. This type of home may be motivated by external recognition or autonomy, competence, and relatedness elements.
Some of the benefits of local motivation inside the home are that the end-user finds rewarding performing activities or changes if they receive outer recognition from the energy community or achieve the reduction with no outer recognition. Additionally, this user can help the community by sharing tips on modifying their habits without affecting, for instance, thermal comfort. Hence, the energy community motivation relies on social sharing and social belonging; the more the user is involved in social sharing and social activities, the more they want to improve and help the others [
37,
49].
Table 6 and
Table 7 show the neuro-fuzzy logic inference rules from the ANFIS system for the summer and winter seasons. The gamification motivation depends on the level of the kWh and setpoint of the house.
Figure 13 shows the summer season rules during weekdays and weekends, and
Figure 14 for the winter season. For Case 1, an interface oriented more to intrinsic motivations is required with a bit of extrinsic motivation during weekends (see
Figure 13a,b). Case 2 requires an interface oriented to the intrinsic motivations(See
Figure 13c,d); on the opposite, Case 3 requires an interface oriented to the extrinsic motivation (See
Figure 13e,f). For the winter periods, Case 1 requires an interface more oriented to extrinsic motivations and a few elements of intrinsic motivation (see
Figure 14a,b) and Case 2 remains with intrinsic motivations and Case 3 with extrinsic motivations (see
Figure 14c–f).
Based on the neuro-fuzzy rules for both periods,
Figure 15 shows the interfaces for each case. The three types of user were selected based on their energy awareness and motivation to modify their energy consumption by changing the time of use of the household appliances:
Case 1. Home-focused—this user is interested in home-improvement efforts while helping to save energy.
Case 2. Green advocate—this user is concerned about saving energy as much as possible.
Case 3. Disengaged energy saver—This user is not interested in saving energy.
The description of gamification elements based on the extrinsic and intrinsic motivation are described as follows:
Case 1. Home-focused energy saver—The intrinsic elements used in
Figure 15a are the notifications, tips, energy community, collaboration, control over peers through competition and social comparison, and the extrinsic elements consider challenges, bill discounts, the levels, and rewards. Moreover, Case 1 is fascinating as this type of user requires a more dynamic interface that changes toward the season and promotes this energy reduction; in that sense, an EMS is ideal for this user type.
Figure 15b displays an interface that places more emphasis on rewards and leader-board elements.
Case 2. Green advocate (
Figure 15c,d)—This interface is oriented more to intrinsic elements, such as social comparison, notifications, tips, energy community, collaboration, control over peers, social comparison, and competition.
Case 3. Disengaged energy saver (
Figure 15e,f)—On the contrary, this interface is oriented to extrinsic elements such as coupons, bill discounts, and challenges. Moreover, a message is displayed as an intrinsic motivator, and this message is focused on showing the end-user the benefits of reducing energy.
In addition to this,
Figure 16a–f show the electrical consumption available in the section of statistics. This electrical consumption allows the user to know how much energy they are using and how they can save energy if they want to do it.
Figure 16g,h display Case 1 for the heater and AC. The interface displays a message connected with the EMS, so several strategies can be used based on the decision tree from
Figure 7.
The EMS automatizes flexible electrical appliances to perform at low or mid electric rates, reduces energy consumption, and guides the user to reduce non-flexible appliances. For example, the water pump, washing machine, and dryer can be used when a low electrical tariff is current in early and late hours to save money. Moreover, the EMS may reduce lighting consumption by sensing user activity in rooms.
5. Discussion
The state of the art reflects the need to combine an EMS with gamification techniques to promote energy reduction. Current frameworks assume a certain level of comfort without considering the user’s preferences and thermal comfort. Moreover, a friendly EMS that displays energy consumption and has auto-configuration or easy set-up is needed to engage the user and optimize consumption when the price is high; in the end, this can help users to reduce electrical consumption [
26,
27,
28,
29,
30,
31,
32]. Thus, considering the human factor while designing an EMS is crucial.
Gamification techniques could help by knowing the types of end-user and proposing specific targets so the users could be engaged. A manner of classifying the type of user has been proposed in [
49], where based on the type of user, tailored gamified interfaces are proposed. Moreover, this paper proposes five user classifications based on the user’s targets, such as saving-money goal, comfort level, environmental commitment, and technology knowledge.
One of the great advantages of AI is the possibility of considering sensors and monitoring the end-user to analyze the level of engagement and determine if, for instance, the gamification elements in the interface are accurate or if it requires changes [
54]. Mainly, this proposal uses the ANFIS decision system to determine which type of gamified motivation is needed to engage the end-user and promote flexible loads during the day.
Although this paper does not consider the inclusion of multi-sensor systems, this could be included in further research for monitoring and tracking the end-user to determine their level of comfort and promote load flexibility based on the users’ daily tasks.
One of the disadvantages of this proposal is the numerous steps required for determining the type of interfaces based on the user; an optimized interface could tackle this disadvantage by providing a previous survey to the user so that the interface could be updated based on their requirements. The simulation could also include longer than one-year historical consumption to better determine the users’ patterns and their thermal comfort depending on the seasons.
Not all the houses or buildings can be used for deploying this technology. The conditions and limits that require a smart home for being benefited of this proposal are connectivity scheme among electrical devices and the monitoring system, in addition to a certain level of control of the flexible ones. The system runs on a smart device such as a cellphone, but it is not the only device that can receive and transmit energy information. On the other hand, energy companies can use the data generated for improving services or facilitate green energy inclusion and stability of the electric grid.
However, this proposal’s advantage is the inclusion of EMS with a gamification structure to provide goal-oriented lucid interfaces, in this case, with the reduction of electrical consumption during peak hours and promoting flexible usage.
6. Conclusions
A gamification strategy and EMS were used to help improve energy efficiency, save energy and money, avoid peak rates, and reduce energy consumption. As a result, this proposal studies energy scenarios with the same energy load schemes (flexible schedule loads and non-flexible loads). Still, for different types of users (user willing to change its consumption patterns without restrictions, user partially willing to modify patterns, and user not flexible), the simulation showed an approximate 10% energy consumption reduction. Aside from AI techniques, fuzzy logic and decision tree were used for the decision-making process, which matched the load scheme and user preferences in a compound tailored interface with the required gamification elements to save energy according to users’ personalities. According to the decision tree system, the fuzzy logic scheme delimits the user preferences to manage the flexible loads (an HVAC system for this case study). Thus, the ANFIS system reaches the tailored interface compound of gamification elements for the rest of the load scheme management for energy efficiency.
Moreover, simulation allows a better decision-making process and forecasting, saving energy and money by making proper use of electrical devices and achieving user goals and preferences. Although this simulation is for consumption per hour and monthly rate, the algorithm allows testing different custom load schemes, dynamic price schemes, and different user behavior. In addition, classifying the type of consumer allows a more accurate profile that helps make decisions required for proposing changes in household appliances. For instance, disengaged energy saver users are not interested in saving energy, so the interface displays gamified extrinsic motivations that motivate them to perform activities to receive rewards. Those activities include the change of thermostat setpoint or the change in household appliance usage during off-peak periods.
On the other hand, green advocate users require interfaces with intrinsic motivations that allow these consumers’ interaction with other types of consumers such as the disengaged or the home-focused, promoting social commitment and social sharing. Therefore, these users can feel part of the community as they help other users reduce energy or promote flexible loads. Further work could provide feedback and adjust the model based on energy consumption to evaluate the overall performance and adapt the interface and the gamification elements. Another aspect to include in evaluating usability and heuristics is to optimize the interface and make it more appealing. In addition, the classification of the user type, their personality traits, and type of gamified user could be included to improve the game dynamics during the application usage.